You are on page 1of 52

194

Chapter-7

Reduced Common Mode Voltage PWM


Algorithms for Direct Torque Controlled
Induction Motor Drives

7.1 Introduction:

Though the SVPWM algorithm reduces harmonic distortion in line

current of the induction motor, it generates large common mode voltage

variations, which is also known as electromagnetic interference (EMI).

The PWM algorithm requires high speed switching devices like IGBTs to

increase the switching frequency of VSI-PWM inverters, which in turns

leads to better performance characteristics. However, the high speed of

variation of the voltage (dv/dt) exerted by such power devices produces

high frequency oscillatory common mode voltages (common mode

currents) at every switching instant.

These oscillatory currents radiate electromagnetic interference

(EMI) noises throughout producing a bad effect on electronic devices

such as amplitude modulation (AM) radio receivers, medical equipments

communication systems. Steep voltage pulses also pose threats to the

motor. Because of the transmission line effect in the motor cable, voltage

and current waves propagate at a finite velocity in the cable, and reflect

from the impedance discontinuities at the motor terminals and the


195

converter output. A superposition of the reflected voltages at the motor

terminals may result in over voltages possibly causing a failure in the

stator winding insulation of the motor. A common-mode voltage is

present in the output voltage of the frequency converter. This voltage

may cause bearing currents in the motor. Bearing currents generate

bearing wear in the form of pitting, frosting, and spark tracks at the

surfaces of the bearing balls and races. The bearings may fail

prematurely because of the additional friction and the polluted lubricant

in the bearing.

Moreover, an increase in the carrier frequency of PWM inverters

results in a non negligible amount of high frequency leakage current

which may cause a serious problem. It would flow through stray

capacitors between stator windings and a motor frame due to a large step

change of the common mode voltage produced by a PWM inverter. The

peak value may reach the rated current in the worst case. It may have an

undesirable influence on the motor current control and may result in

incorrect operation of residual current operated circuit breakers. Many

studies for reducing the common mode voltage have been made. These

studies, however, have focused on the utilization of additional hardware

such as various types of inverter output filters and an active cancellation

using switching devices. Since these methods require extra hardware, the

drawbacks of an increase in inverter weight and volume or complexity in

its control are unavoidable.


196

Recently various common mode voltage reduction methods using a

PWM scheme have been proposed in the literature. The PWM schemes

are based on space vector approach and can be of symmetrical or

asymmetrical. Moreover, they do not use any zero voltage vectors which

are cause of high common mode voltage. In case of asymmetrical PWM

technique the switching number inside the sector is not the same and

switching pattern is not symmetrical as in conventional SVPWM method

and hence increases current disturbance. In case of symmetrical space

vector based PWM technique though the switching pattern is

symmetrical and number of switchings within the sector remains the

same as in conventional SVPWM, but the number of switchings from one

sector to the next is more and hence switching losses increases.

In this chapter, few simplified space vector based PWM algorithms

have been proposed for direct torque controlled induction motor drives to

reduce the common mode voltage. These PWM algorithms are also known

as reduced common mode voltage PWM (RCMVPWM) algorithms. As the

proposed RCMVPWM algorithms are developed based on the concept of

imaginary switching times, the computational burden involved in the

algorithms can be decreased.

7.2 Common Mode Voltage:

The common mode voltage is the potential of the star point of the

load with respect to the center of the dc bus of the VSI. Fig 7.1 shows a
197

VSI fed induction motor. A set of phase voltage equations can be written

as given in (7.1).

Van = V so Vao
Vbn = V so Vbo (7.1)
Vcn = V so Vco

where Vao , Vbo , Vco are inverter pole voltages and V so is common mode

voltage.

3-ph, 50 Hz Induction Motor


AC supply Zn
Va
o Vb s
V dc
Vc

V so

Fig 7.1 Three-phase VSI fed induction motor

Adding the set of equations of (7.1) and sinceVan + Vbn + Vcn = 0 , the

common mode voltage in the motor is given by

V + Vbo + Vco
Vcom = V so = ao (7.2)
3

Hence, if the drive is fed by balanced three phase supply, the common

mode voltage is zero. But, the common mode voltage exists inevitably

when the drive is fed from an inverter employing PWM technique because

the VSI cannot produce pure sinusoidal voltages and has discrete output

voltages. It can be shown that the switching state and dc bus voltage
198

decides the common mode voltage. There are eight available output

voltage vectors in accordance with the eight different switching states of

the inverter as depicted in Fig. 7. 2.

V3 (- + -) V2 (+ + -)

V4 (- ++) V7 (+ ++)
V1 (+ - -)
V0 (- - -)

V5 (- - +) V6 (+ - +)

Fig 7.2 Voltage space vectors of a three-phase VSI

According to the switching states of the inverter the common mode

voltage can be expressed as given in (7.3).

V V
Vcom = V so = dc (S a + S b + S c ) dc (7.3)
3 2

where Sa, Sb and Sc denotes the switching states of each phase.

The common mode voltage for each inverter state is given in Table

7.1, which shows that, if only even or only odd voltage vectors are used,

no common mode voltage variation is generated. If a transition occurs

from an even voltage vector to an odd one (or vice versa), a common

mode variation of amplitude Vdc/3 is generated. If a transition from an

odd (even) voltage vector to the zero (seventh) voltage vector occurs, a
199

common variation Vdc/3 is generated. If a transition from an odd (even)

voltage vector to the seventh (zero) voltage occurs, a common-mode

variation of amplitude 2Vdc/3 is generated. Finally, if a transition occurs

from zero to seventh or vice versa, a common mode variation of

amplitude Vdc is generated.

Table 7.1 Common mode voltage and output voltage generated by a


switching state
Switching
state Inverter pole voltages
V com

Vao Vbo Vco


V 0 (0 0 0) V V V V
dc dc dc dc
2 2 2 2
V1 (1 0 0) V dc V dc V dc V dc

2 2 2 6
V 2 (1 1 0) V dc V dc V dc V dc

2 2 2 6
V 3 (0 1 0) V V dc V V
dc dc dc
2 2 2 6
V 4 (0 1 1) V V dc V dc V dc
dc
2 2 2 6
V 5 (0 0 1) V V Vdc V
dc dc dc
2 2 2 6
V 6 (1 0 1) V dc V V dc V dc
dc
2 2 2 6
V 7 (1 1 1) V dc V dc V dc V dc
2 2 2 2

Therefore, from the point of view of common mode emissions, the

worst case is transition between two zero voltage vectors. For this reason

whatever inverter control technique is devised, to minimize the generated

common mode emissions of the drive, the exploitation of both the null

voltage vectors (zero and seventh) should be avoided.


200

7.3 Reduced Common Mode Voltage PWM Algorithms:

In the application field the problems related to common mode

voltage are increasing due to increased PWM switching frequencies aimed

at higher efficiency, increased bandwidth, etc. Hence, common mode

voltage reduction techniques have been gaining importance. The effect of

common mode voltage can be reduced actively or passively. The active

common mode voltage reduction method that involves controlling the

PWM pulse patterns is the most economical method as it requires no

extra components. Recently, several PWM pulse patterns that yield

reduced common mode voltage termed as reduced common mode voltage

PWM (RCMVPWM) methods have been reported [93-96]. In all these

methods, the zero states of the inverter are avoided and results in a

common mode voltage of Vdc/6. All these methods are described using

space vector approach.

Based on the choice of the voltage vectors, the RCMVPWM methods

will be sub grouped in three types, these include active zero state PWM

(AZSPWM) algorithms, remote state PWM (RSPWM) algorithms and near

state PWM (NSPWM) algorithm. In the AZSPWM algorithms, the

conventional active (adjacent) voltage vectors are complemented with

either two near opposing active vectors or one of the adjacent states and

its opposite vector with equal time to effectively create a zero-voltage

vector. The RSPWM methods synthesize the output voltage from three

inverter voltage vectors that are 120o apart from each other (most remote
201

vectors). The NSPWM method utilizes a group of three neighbor voltage

vectors to match the output and reference volt-seconds. These three

voltage vectors are selected such that the voltage vector closest to

reference voltage vector and its two neighbors (to the right and left) are

utilized. But, this thesis mainly deals with only AZSPWM and NSPWM

algorithms.

In the AZSPWM algorithms, the vector transformation yields six

active and two zero vectors for the inverter, and as shown in Fig. 7.3 (a),

the vectors divide the space into six segments as in SVPWM algorithm.

These regions are utilized in programming the PWM pulses. But, in the

NSPWM algorithm 30o phase-shifted regions are utilized as shown in Fig.

7.3 (b).

In the space vector approach, the duty cycles of the voltage vectors

are calculated according to the vector volt-seconds balance rule. The

voltage vectors and their sequences are selected based on a specified

performance criterion such as the minimum output voltage ripple and

switching count and the vectors are programmed accordingly. The

formation of AZPWM and NSPWM algorithms is illustrated in Fig. 7.4

Fig. 7.5. In each PWM method, with a specific performance optimization

criterion, the voltage vectors are selected, and their sequences depend on

the region of the reference voltage vectors defined in Fig. 7.3(a) and (b).

The utilized voltage vectors and their sequences for the conventional

SVPWM and RCMVPWM methods are given in Table 7.2.


202

V3 V2

A2

A3 A1
V4 V7 V1
V0
A4 A6
A5

V5 V6
(a)

V3 V2

B2
B3

V4 B4 V7 B1
V1
V0

B5 B6

V5 V6
(b)
Fig. 7.3 Voltage space vectors and 60o sector definitions

V3 V2

A2

A3 A1
V4 V7 V1
V0
A4 A6
A5

V5 V6
Fig. 7.4 Voltage space vectors and formation of AZSPWM algorithms
203

V3 V2

B2
B3

V4 B4 V7 B1
V1
V0

B5 B6

V5 V6
Fig. 7.5 Voltage space vectors and formation of NSPWM algorithm

Table 7.2 Region dependent pulse patterns of various PWM


techniques
PWM
A1 A2 A3 A4 A5 A6
Technique
Conventional 0127- 0327- 0347- 0547- 0567- 0167-
SVPWM 7210 7230 7430 7450 7650 7610
3216- 4321- 5432- 6543- 1654- 2165-
AZSPWM1
6123 1234 2345 3456 4561 5612
5122- 6233- 1344- 2455- 3566- 4611-
AZSPWM2
2215 3326 4431 5542 6653 1164
B1 B2 B3 B4 B5 B6

216- 321- 432- 543- 654- 165-


NSPWM
612 123 234 345 456 561

In the conventional SVPWM method, two adjacent states with two

zero voltage vectors are utilized to program the output voltage. Every 600

degrees the active voltage vectors change, but the zero state locations are

retained. In the AZSPWM methods, the choice and the sequence of active

voltage vectors are the same as in conventional SVPWM. However,

instead of the real zero voltage vectors (V0 and V7), two active opposite

voltage vectors with equal duration are utilized. Here, three choices exist.

For AZSPWM algorithms, any of the pairs V1-V4, V2-V5, or V3-V6 can be
204

utilized.

The NSPWM employs only three neighbor voltage vectors and

sequences them in the order that the minimum switching count is

obtained. Thus, one of the phases is not switched within each PWM

cycle. The resulting vector sequence is shown in Table 7.2. For example,

for the region between -30o and +30o (sector B1), the applied voltage

vectors are V1, V2, and V6 with the sequence V6-V2-V1-V1-V2-V6. In the

NSPWM algorithm, in each sector any one of the phases is clamped to

either positive or negative DC bus for a total of 120o over a fundamental

cycle. Hence, it reduces the switching losses of the inverter and switching

frequency of the inverter by 33.33%.

For each one of the above described PWM methods and its pulse

pattern, a unique common mode voltage pattern results. The

conventional SVPWM method has common mode voltage (CMV) that can

be as high as Vdc/2. On the other hand, all the RCMVPWM methods

have Vdc/6 CMV magnitude value. The frequency and polarity of CMV

of each RCMVPWM method is unique and each of these pulse patterns

may be attractive for a specific application. All the above existing

RCMVPWM algorithms use the conventional space vector approach,

which requires angle and sector information. Hence, to reduce the

complexity involved in the existing RCMVPWM algorithms, the proposed

RCMVPWM algorithms have been proposed by using the concept of

imaginary switching times.


205

7.3.1 Calculation of Switching Times for AZSPWM Algorithms:

To reduce the complexity involved in the existing AZSPWM

algorithms, in this chapter, the proposed AZSPWM have been developed

by using the notion of imaginary switching times. The concept of

imaginary switching times has already explained in chapter 3 in detail.

In this approach, the imaginary switching time periods proportional to

the instantaneous values of the reference phase voltages are calculated

as given in (7.4)

T T T
Tan s Van ; Tbn s Vbn ; Tcn s Vcn (7.4)
Vdc Vdc Vdc

To calculate the active vector switching times, the maximum, middle and

minimum values of imaginary switching times are calculated in every

sampling time as given in (7.5) (7.7).

Tmax = Max (Tan , Tbn , Tcn ) (7.5)

Tmid = Mid(Tan , Tbn , Tcn ) (7.6)

Tmin = Min (Tan , Tbn , Tcn ) (7.7)

Then the active vector switching times T1 and T2 may be expressed as

T1 = T max Tmid ; T2 = Tmid T min (7.8)

The zero voltage vectors switching time is calculated as

Tz = Ts T1 T2 (7.9)

Then, the zero voltage vector time is shared equally among the two

opposite active voltage vectors to create effectively a zero voltage vector.

Then, by utilizing the space vector approach, the possible switching


206

sequences, can be derived. However, this thesis presents two cases oz

AZSPWM algorithms for which the sequences are illustrated in Table 7.2.

7.3.2 Generation of Pulse Pattern for NSPWM Algorithm:

7.3.2.1 Conventional Approach:

The near state PWM (NSPWM) algorithm uses a group of three

neighbor voltage vectors to construct the reference voltage vector. In

order to reduce the common mode voltage variations, the proposed

NSPWM algorithm did not use the zero voltage vectors. These three

voltage vectors are selected such that the voltage vector closest to

reference voltage vector and its two neighbors are utilized in each sector

as shown in Fig 7.5. Hence, the utilized voltage vectors are changed in

every sector. As shown in Fig. 7.5, to apply the method, the voltage

vector space is divided into six sectors. Here also, as all six sectors are

symmetrical, the discussion is limited to the first sector only. For the

required reference voltage vector, the active voltage vectors (V1, V2 and

V6) times can be calculated as in (7.10), (7.11) and (7.12) [93-96].

3 3 3
T1 = 1 + M i cos( + ) + M i sin( + )Ts (7.10)
3 3

3 3
T2 = 1 M i cos( + ) M i sin( + )Ts (7.11)
3 3

T6 = Ts T1 T2 (7.12)

But, in the NSPWM algorithm, the (7.10), (7.11) and (7.12) have a valid

solution when the modulation index is varying between 0.61 and 0.906
207

[93-96]. In order to get minimum switching frequency and reduced

common mode voltage the NSPWM algorithm uses 216-612 in sector-I,

321-123 in sector-II and so on.

7.3.2.2 Proposed Approach:

As the existing NSPWM algorithm uses conventional space vector

approach, the complexity involved in the algorithm is more. To simplify

the algorithm, proposed NSPWM algorithm is developed by using the

notation of imaginary switching times. The modulating waveform of

NSPWM algorithm is similar to the discontinuous PWM 1 (DPWM1)

algorithm as explained in [93-96]. The modulating waveform can be

generated by using the concept of imaginary switching times as given

below:

The imaginary switching time periods, which are proportional to

the instantaneous values of the reference phase voltages, are calculated

from (7.4). Then the maximum and minimum values can be calculated

by using (7.5) and (7.7). Then, the effective time during which the

induction motor is effectively connected to the source (that is the power

will be transferred to the motor from source) can be calculated as given

in (7.13).

Teff = Tmax Tmin (7.13)

When the actual gating signals for power devices are generated in

the PWM algorithm, there is one degree of freedom by which the effective

time can be relocated anywhere within the sampling time period.


208

Therefore, the actual switching times for each inverter leg can be

obtained by the time shifting operation as follows:

Tga = Tan + Toffset (7.14)

Tgb = Tbn + Toffset (7.15)

Tgc = Tcn + Toffset (7.16)

To guarantee the full utilization of dc-link voltage of the inverter,

the actual switching times should be restricted to a value from 0 to Ts .

To generate the modulating waveforms of NSPWM algorithm, the

procedure is as follows:

If the A phase reference voltage is positive (or negative) and has

maximum magnitude, the A-phase switch should be fixed to the ON (or

OFF) state. That is to say,

if T max + T min, < 0 T max + Toffset = Ts


(7.17)
if T max + T min 0 T min + Toffset = 0

Therefore, the time shifting value Toffset is

if T max + T min, < 0 Toffset = T min


(7.18)
if T max + T min 0 Toffset = Ts T max

Then, the modulating waveforms of NSPWM algorithm can be

synthesized using the calculated gating times by using (7.19).

V 2 * T gi
*
Vin = dc 1 i = a , b, c (7.19)
2 Ts

The modulating waveforms of NSPWM and DPWM1 algorithm are


209

exactly the same [93-96] and Fig 7.6 shows the waveforms of offset time,

gating time and modulating signal of NSPWM algorithm.

(a)

(b)
Fig 7.6 waveforms of (a) actual gating time (Tga) and offset time
(Toffset) (b) Modulating wave

The total number of commutations in SVPWM algorithm is three in

a sampling time interval, where as the number of commutations in

NSPWM algorithm is two. Moreover, from the modulating waveform of

NSPWM algorithm, it can be observed that any one of the phases is

clamped to the either positive or negative DC bus for utmost a total of


210

1200 over a fundamental cycle. Hence, the switching losses of the

associated inverter leg are eliminated. Hence, the switching frequency of

the NSPWM algorithms is reduced by 33% compared with SVPWM

algorithm. This implies that if one wants to keep the average switching

frequency constant, whenever NSPWM algorithm is used, sampling

3
frequency must be increased by times. However, in NSPWM
2

algorithm instead of one carrier wave, two carrier waves (Vtri and Vtri)

must be utilized. The choice of the triangle to be compared with the

modulation signals is region dependent. If slope of the reference phase

voltage is positive then the modulating waveform is compared with Vtri

and if slope of the reference phase voltage is negative then the

modulating waveform is compared with -Vtri. General switching rule is

that if the modulating waveform is larger than the carrier signal

(triangular wave), the upper switch associated with the specific phase is

set on. The reference phase voltages are as shown in Fig 7.7 in all the

six sectors.

From Fig 7.7, it can be observed that in the first sector the slope of

a- and b-phase voltages is positive, whereas the slope of c-phase voltage

is negative. Hence, in the first sector, the modulating waveforms of a-

phase and b-phase are compared with Vtri and modulating wave of c-

phase is compared with Vtri as shown in Fig 7.8. The possible pulse

pattern for the first sector is also shown in Fig 7.8.


211

1 Va Vb

Vc

Voltages 0

-1
0 60 120 180 240 300 360
Angle (degrees)
Fig 7.7 reference phase voltages in all sectors in NSPWM algorithm
Va*

-Vtri
+Vtri
Vb*

Vc*

Sa

Sb

Sc
2 1 6 6 1 2

Fig 7.8 Scalar implementation of NSPWM algorithm (in first sector)

7.4 Proposed RCMVPWM Algorithms Based Direct Torque Controlled

Drive:

The reference voltage space vector can be constructed in many

ways. But, to reduce the complexity of the algorithm, in this thesis, the

required reference voltage vector, to control the torque and flux cycle-by-

cycle basis is constructed by using the errors between the reference d-


212

axis and q-axis stator fluxes and d-axis and q-axis estimated stator

fluxes sampled from the previous cycle. The block diagram of the

proposed RCMVPWM based DTC is as shown in Fig. 7.9. From Fig. 7.9, it

is seen that the proposed RCMVPWM based DTC scheme retains all the

advantages of the CDTC, such as no co-ordinate transformation, robust

to motor parameters, etc. However a space vector modulator is used to

generate the pulses for the inverter, therefore the complexity is increased

in comparison with the CDTC method.

In the proposed method, the position of the reference stator flux

vector s * is derived by the addition of slip speed and actual rotor speed.

The actual synchronous speed of the stator flux vector s is calculated

from the adaptive motor model.

After each sampling interval, actual stator flux vector s is

corrected by the error and it tries to attain the reference flux space

vector s * . Thus the flux error is minimized in each sampling interval.

The d-axis and q-axis components of the reference voltage vector can be

obtained as follows:

Reference values of the d-axis and q- axis stator fluxes and actual

values of the d-axis and q-axis stator fluxes are compared in the

reference voltage vector calculator block and hence the errors in the d-

axis and q-axis stator flux vectors are obtained as in (7.20)-(7.21).


213

Vdc

Te sl e e V ds
Reference R
SMC
+
PI
+ Voltage C
Ref M
_ Vector
speed Te r + Calculator
V
P
+ s* W
_ V qs* M
Speed
Vds, Vqs
s Calculation

Adaptive
Motor Model 3
2

IM

Fig. 7.9 Block diagram of proposed RCMVPWM based DTC.

ds = ds * ds (7.20)

qs = qs * qs (7.21)

The knowledge of flux error and stator ohmic drop allows the

determination of appropriate reference voltage space vectors as given in

(7.22)-(7.23).

* = ds
V ds R s i ds + (7.22)
Ts

qs
* = (7.23)
V qs R s i qs +
Ts

Where, Ts is the duration of subcycle or sampling period and it is a half of

period of the switching frequency. This implies that the torque and flux

are controlled twice per switching cycle. Further, these d-q components
214

of the reference voltage vector are fed to the SVPWM block from which,

the actual switching times for each inverter leg are calculated.

7.5 Simulation Results and Discussions:

To verify the proposed algorithms, numerical simulation studies

have been carried out on a v/f controlled induction motor drive by using

Matlab/Simulink. For the simulation, the average switching frequency of

the inverter is taken as 3 kHz. The parameters of the induction motor are

given in Appendix - I.

The simulation results of common mode voltage variations for

SVPWM and RCMVPWM algorithms are given from Fig.7.10 to Fig.7.13.

From Fig.7.10 it can be observed that as the SVPWM algorithm uses two

zero voltage vectors i.e, V 0 (000) and V7 (111) in each sampling time

interval, the common mode voltage varies from Vdc 2 to + Vdc 2 .

Hence, to reduce the common mode voltage variations, AZSPWM and

Fig.7.10 Variation of common mode voltage for SVPWM


215

Fig.7.11 Variation of common mode voltage for AZSPWM1

NSPWM algorithms have been proposed in this chapter. As the proposed

AZSPWM and NSPWM algorithms do not use the zero voltage vectors, the

common mode voltage will vary from + Vdc 6 to Vdc 6 as shown in Fig.

7.11 Fig. 7.13.

Fig.7.12 Variation of common mode voltage for AZSPWM2


216

Fig. 7.13 Variation of common mode voltage for NSPWM

The simulation results of SVPWM and RCMVPWM based v/f

controlled induction motor drive are shown in Fig 7.14 - Fig. 7.17. The

harmonic spectra of line currents are also shown in Fig. 7.14 - Fig. 7.17.

From the simulation results, it can be observed that the SVPWM

algorithm gives less harmonic distortion with more common mode

voltage variations. Whereas, the RCMVPWM algorithms (both AZSPWM

and NSPWM) will give reduced common mode voltage variations with

increased harmonic distortion.


217

(a)

(b)

Fig. 7.14 (a) Steady state plots (b) Harmonic spectra of line current

for SVPWM at Mi=0.815 (f=45Hz)


218

(a)

(b)

Fig. 7.15 (a) Steady state plots (b) Harmonic spectra of line current

for AZSPWM1 at Mi=0.815 (f=45Hz)


219

(a)

(b)

Fig.7.16 (a) Steady state plots (b) Harmonic spectra of line current

for AZSPWM3 at Mi=0.815 (f=45Hz)


220

(a)

(b)

Fig.7.17 (a) Steady state plots (b) Harmonic spectra of line current

for NSPWM at Mi=0.815 (f=45Hz)


221

The simulation results of proposed AZSPWM1 algorithm based

direct torque controlled induction motor drive are shown from Fig 7.18 to

Fig 7.29. Fig 7.18 shows the starting transients of speed, torque stator

currents and flux in proposed AZSPWM1 algorithm based drive, from

which it can be observed that the AZSPWM1 algorithm gives more ripple

in starting torque when compared with the SVPWM algorithm. Fig 7.18

shows the no-load starting transients of speed, currents, torque and flux.

Fig. 7.19 shows the starting transients in phase and line voltages for

AZSPWM1 algorithm based DTC-IM drive. The no-load steady state plots

of speed, torque, stator currents, flux, phase and line voltages at 1200

rpm are given in Fig 7.20 -Fig. 7.21. The harmonic distortion in the

steady state stator current along with THD value is shown in Fig 7.22.

From Fig 7.20 to Fig 7.22, it can be observed that the steady state ripple

in torque, flux and current is increased when compared with SVPWM

based DTC. The locus of the stator flux is given in Fig 7.23, from which it

can be observed that the locus is almost is a circle of constant radius.

The transients in speed, torque, currents and flux during the step

change in load torque and corresponding phase and line voltages are

shown in Fig. 7.24 -Fig. 7.25. Also, the transients in speed, torque,

currents, flux, and voltages during the speed reversals (from +1200 rpm

to -1200 rpm and from -1200 rpm to +1200 rpm) are shown from Fig.

7.26 to Fig. 7.29. The four-quadrant speed-torque characteristic of the

proposed drive is shown in Fig. 7.30.


222

Fig. 7.18 Simulation results of AZSPWM1 based DTC: starting


transients.

Fig. 7.19 Starting transients in phase and line voltages for


AZSPWM1 algorithm based DTC drive.
223

Fig. 7.20 Simulation results of AZSPWM1 based DTC: steady-state


plots at 1200 rpm.

Fig. 7.21 The phase and line voltages of AZSPWM1 based DTC drive
during the steady state operation.
224

Fig. 7.22 Harmonic Spectrum of stator current along with THD for
AZSPWM1 based DTC-IM drive.

Fig. 7.23 locus of stator flux in AZSPWM1 based DTC-IM drive.


225

Fig. 7.24 Transients during step change in load torque (a 30 N-m


load torque is applied at 0.5 s and removed at 0.6s).

Fig. 7.25 The phase and line voltages of AZSPWM1 based DTC during
step change in load torque (a 30 N-m load torque is applied at 0.5 s
and removed at 0.6s).
226

Fig. 7.26 Transients in speed, torque, current and flux during speed
reversal (speed is changed from +1200 rpm to -1200 rpm at 0.7 s).

Fig. 7.27 Transients in phase and line voltages during speed reversal
(speed is changed from +1200 rpm to -1200 rpm at 0.7 s).
227

Fig. 7.28 Transients in speed, torque, current and flux during speed
reversal (speed is changed from -1200 rpm to +1200 rpm at 1.35 s).

Fig. 7.29 Transients in phase and line voltages during speed reversal
(speed is changed from -1200 rpm to 1200 rpm at 1.35 s).
228

Fig. 7.30 Speed-torque characteristic of AZSPWM1 based DTC drive


in four-quadrants.

Fig. 7.31 simulation results of AZSPWM2 based DTC: starting


transients.
229

Fig 7.31 shows the no-load starting transients of speed, currents,

torque and flux. Fig. 7.32 shows the starting transients in phase and line

voltages for AZSPWM2 algorithm based DTC-IM drive.

Fig. 7.32 Starting transients in phase and line voltages for


AZSPWM2 algorithm based DTC drive.

The no-load steady state plots of speed, torque, stator currents,

flux, phase and line voltages at 1200 rpm are given in Fig 7.33 - Fig.7.34.

The harmonic distortion in the steady state stator current along with

THD value is shown in Fig 7.35. From Fig 7.31 to Fig 7.35, it can be

observed that the steady state ripple in torque, flux and current is more

when compared with SVPWM based DTC. The locus of the stator flux is

given in Fig 7.36, from which it can be observed that the locus is almost

is a circle of constant radius.


230

Fig. 7.33 Simulation results of AZSPWM2 based DTC: steady-state


plots at 1200 rpm.

Fig. 7.34 the phase and line voltages of AZSPWM2 based DTC drive
during the steady state operation.
231

Fig. 7.35 Harmonic Spectrum of stator current along with THD for
AZSPWM2 based DTC-IM drive.

Fig. 7.36 locus of stator flux in AZSPWM2 based DTC-IM drive.

The transients in speed, torque, currents and flux during the step

change in load torque and corresponding phase and line voltages are

shown in Fig. 7.37 - Fig. 7.38. Also, the transients in speed, torque,

currents, flux, and voltages during the speed reversals (from +1200 rpm
232

Fig. 7.37 Transients during step change in load torque (a 30 N-m


load torque is applied at 0.5 s and removed at 0.6s).

Fig. 7.38 The phase and line voltages of AZSPWM1 based DTC during
step change in load torque (a 30 N-m load torque is applied at 0.5 s
and removed at 0.6s).
233

Fig. 7.39 Transients in speed, torque, current and flux during speed
reversal (speed is changed from +1200 rpm to -1200 rpm at 0.7 s).

Fig. 7.40 Transients in phase and line voltages during speed reversal
(speed is changed from +1200 rpm to -1200 rpm at 0.7 s).
234

Fig. 7.41 Transients in speed, torque, current and flux during speed
reversal (speed is changed from -1200 rpm to +1200 rpm at 1.35 s).

Fig. 7.42 Transients in phase and line voltages during speed reversal
(speed is changed from -1200 rpm to 1200 rpm at 1.35 s).
235

Fig. 7.43 Speed-torque characteristic of AZSPWM2 based DTC drive


in four-quadrants.

to -1200 rpm and from -1200 rpm to +1200 rpm) are shown from Fig.

7.39 to Fig. 7.42. The four-quadrant speed-torque characteristic of the

proposed drive is shown in Fig. 7.43.

Fig 7.44 shows the no-load starting transients of speed, currents,

torque and flux. Fig. 7.45 shows the starting transients in phase and line

voltages for NSPWM algorithm based DTC-IM drive. The no-load steady

state plots of speed, torque, stator currents, flux, phase and line voltages

at 1200 rpm are given in Fig. 7.46 - Fig. 7.47. The harmonic distortion in

the steady state stator current along with THD value is shown in Fig

7.48. From Fig 7.44 to Fig 7.48, it can be observed that the steady state

ripple in torque, flux and current is less compared with AZSPWM based

DTC and slightly more when compare with the SVPWM based DTC.
236

Fig. 7.44 simulation results of NSPWM based DTC: starting


transients.

Fig. 7.45 starting transients in phase and line voltages for NSPWM
algorithm based DTC drive.
237

Fig. 7.46 Simulation results of NSPWM based DTC: steady-state plots


at 1200 rpm.

Fig. 7.47 the phase and line voltages of NSPWM based DTC drive
during the steady state operation.
238

Fig. 7.48 Harmonic Spectrum of stator current along with THD for
NSPWM based DTC-IM drive.

Fig. 7.49 locus of stator flux in NSPWM based DTC-IM drive.

The locus of the stator flux is given in Fig 7.49, from which it can be

observed that the locus is almost is a circle of constant radius. The

transients in speed, torque, currents and flux during the step change in

load torque and corresponding phase and line voltages are shown in Fig.
239

7.50 - Fig. 7.51. Also, the transients in speed, torque, currents, flux, and

voltages during the speed reversals (from +1200 rpm to -1200 rpm and

from -1200 rpm to +1200 rpm) are shown from Fig. 7.52 to Fig. 7.55. The

four-quadrant speed-torque characteristic of the proposed drive is shown

in Fig. 7.56.

Fig. 7.50 Transients during step change in load torque (a 30 N-m


load torque is applied at 0.5 s and removed at 0.6s).
240

Fig. 7.51 The phase and line voltages of NSPWM based DTC during
step change in load torque (a 30 N-m load torque is applied at 0.5 s
and removed at 0.6s).

Fig. 7.52 Transients in speed, torque, current and flux during speed
reversal (speed is changed from +1200 rpm to -1200 rpm at 0.7 s).
241

Fig. 7.53 Transients in phase and line voltages during speed reversal
(speed is changed from +1200 rpm to -1200 rpm at 0.7 s).

Fig. 7.54 Transients in speed, torque, current and flux during speed
reversal (speed is changed from -1200 rpm to +1200 rpm at 1.35 s).
242

Fig. 7.55 Transients in phase and line voltages during speed reversal
(speed is changed from -1200 rpm to 1200 rpm at 1.35 s).

Fig. 7.56 Speed-torque characteristic of NSPWM based DTC drive in


four-quadrants.
243

The common mode voltage variations of SVPWM, AZSPWM and

NSPWM base DTC are shown in Fig. 7.57- Fig. 7.60, from which it can be

observed that the proposed RCMVPWM algorithms will give reduced

common mode voltage variations and hence reduce EMI.

Fig. 7.57 Common mode voltage variations in SVPWM base DTC


drive

Fig. 7.58 Common mode voltage variations in AZSPWM1 base DTC


drive
244

Fig. 7.59 Common mode voltage variations in AZSPWM2 base DTC


drive

Fig. 7.60 Common mode voltage variations in NSPWM base DTC


drive
245

7.6 Summary:

Though the SVPWM based DTC gives less harmonic distortion and

steady state ripple, it gives large common mode voltage variations. This

will cause EMI problems in the drive systems. Hence, to reduce the EMI

problems and to reduce the common mode voltage variations, simplified

RCMVPWM algorithms are presented in this thesis. From the results, it

can be observed that the proposed RCMVPWM algorithms will give

reduce common mode voltage variations with slightly increased ripple

and THD. Moreover, among the presented RCMVPWM algorithms, the

NSPWM algorithm gives superior performance and also it gives less

switching losses of the inverter when compared with the remaining

RCMVPWM algorithms.